Phonetic Patterns after Co2 Laser Posterior Transverse Cordotomy for Bilateral Vocal Fold Paralysis

A combination of non invasive acoustic and aerodynamic measures were used to analyze vocal function in three patients treated by CO2 Laser posterior transverse cordotomy for bilateral vocal fold paralysis. Measures included : (i) Frequency features : fundamental frequency, standard deviation, jitter, shimmer and harmonic-to-noise ratio, (ii) Several measures on the spectra of vowel /a/ : firstformant banwidths, the difference between the amplitudes of the first and second harmonics (H1-H2), the difference between the amplitude of the first harmonic and the amplitude of the first formant peak (H1-A1), the difference between the amplitudes of the first harmonic and the third formant peak was measured (H1-A3). (iii) Laryngeal aerodynamic parameters include maximum phonation time (MPT), mean air flow rate in « sustained phonation » (MFRs), phonation quotient (PQ) and mean air flow rate in « comfortable phonation » (MFRc). These measures are discussed to examine relationships among acoustic and laryngeal aerodynamic parameters, and to evaluate the degree of breathiness. 1INTRODUCTION Surgical management addressed bilateral vocal fold paralysis (BVFP) in patients with severe obstruction, should aim at a compromise between respiratory and phonatory performance. In the past decade, endoscopic CO2 laser arytenoidectomy has become perhaps the most commonly used surgical procedure for enlarging the glottis airway in patients whith BVFP [1] (Figure 2). Endoscopic CO2 Laser posterior transverse cordotomy (PTC) was initially reported by Dennis and Kashima in 1989 [2, 3] (Figure 3). Using the surgical properties of the CO2 Laser, the vocal ligament is sectionned just anterior to the vocal process of the arytenoid cartilages. Dennis and Kashima reported that the CO2 Laser PTC enlarges the posterior glottic aperture « respiratory glottis » and preserves close approximation of the anterior membranous vocal cord for phonation « phonatory glottis » (Figure 1). In retrospective study, we recently reported the safety and efficiency of this technique in 25 patients with BVFP [4]. Many normal subjects configure the vocal folds so that the glottis is never completely closed throughout a cycle of vibration. In normal subjects, « breathy voice » is defined as the modifications introduced by fixed opening between the arytenoid cartilages : (i) an increase in the bandwidth of the first formant, (ii) an increased tilt in the spectrum at high frequencies, and (iii) emergence of a turbulence noise source (glottis) [5-8] The aims of this study were to analyze vocal function and phonetic patterns after CO2 Laser posterior transverse cordotomy (PTC) in bilateral vocal fold paralysis. 2METHODS AND RESULTS Four patients were included in this prospective study and recorded preoperatively and postoperatively at 1, 3, 6, 12 and 24 months. The etiologies included two cases of BVFP following surgical trauma (total thyroidectomy) and two cases of central nervous system disease. Distribution by sex was two male (M1 : 77 years and M2 : 69 years) and two female patients (F1 : 74 years and F2 : 52 years). CO2 Laser PTC was previously described [2, 4]. All patients were recorded under similar conditions. Pretreatment measurements were not available for one patient (M2) because he has been treated in emergencies. 2.1. Methods 2.1.1. Acoustic parameters. Two tasks were completed as part of this project. Patients were recorded in a quiet room and instructed to produce a number of sustained vowels /a/ at habitual conversational pitch and loudness. In the first task, frequency features were analyzed with the Computerized Speech Lab and the Multidimentional Voice Program (Kay Elemetrics, NJ, USA). The frequency features automatically recorded were the average fundamental frequency (FO) in Herz, the FO standard deviation (SD) in Herz, the jitter in percent, the shimmer in percent, and the harmonic-to-noise ratio (HNR) in percent. In the second task, several data were extracted on the spectra of vowels /a/ at 24 months after Laser procedure. Measures for a given patient are averages across five repetitions for each vowel. Before 24 month, instability of the vibration pattern was too higher, and it was not possible to measure data on the spectra of vowel /a/. The following data were extracted. (i) First-formant banwidths, (ii) The difference between the amplitudes of the first and second harmonics (H1-H2). H1-H2 was measured at the same points where F1 bandwidth was estimated. (iii) The difference between the amplitude of the first harmonic and the amplitude of the first formant peak (H1-A1). (iv) The difference between the amplitudes of the first harmonic and the third formant peak was measured (H1-A3). A non parametric Wilcoxon test was used for each comparison among preoperatively and postoperatively for M2, F1 and F2. page 811 ICPhS99 San Francisco 2.1.2. Aerodynamic parameters. Aerodynamic parameters were performed with the Aerophone II (Kay Elemetrics, NJ, USA). Two tasks were completed as part of this project. The patients produced the vowel /a/ for as long as possible after a maximal inspiration, in « maximum sustained phonation », or following comfortable inspiration « comfortable phonation ». The maximum phonation time (MPT), the mean air flow rate (MFRs) and the phonation quotient (PQ) were measured in « maximum sustained phonation ». The PQ is defined as the vital capacity divided by the MPT. The mean air flow rate (MFRc) was determined in « 2comfortable phonation ». For the two tasks, the patients repeated the test several times (minimum three attempts), resting between each attempt. The method was first demontrated by the experimenter. Data from each recordings of the one male and two female patients were compared with the male and female reference group data by calculating Z scores according Holmberg [9]. Group reference data were obtained from 60 male and 50 female patients [10]. The Z score was defined as the ratio of the observed value minus the group mean value divided by the group standard deviation. Z score calculations were based on data for the separate male and female reference groups [10]. These measures are used to examine relationships among laryngeal aerodynamic parameters and acoustic characteristics of voice for M1, F1 and F2. 2.2. Results Frequency features and laryngeal aerodynamic parameters for one male (M1) and two female (F1, F2) patients are presented in Table 1. Measures on the spectra of the vowel /a/, first-formant bandwidth, H1-H2, H1-A1, H1-A3, differed significantly among postoperatively and preoperatively patients (M1, F1 and F2). Frequency features Preoperative 1 and 3 months 6 months 12 and 24 months Average F0 in Herz N ND N N SD F0 in Herz N ND + + Jitter in percent N ND + + Shimmer in percent N ND + + HNR ratio in percent N ND + + Aerodynamic parameters MPT in seconds N + + N MFRs in liters/second N + + N PQ in liters/second N + + N